Photofragment imaging of methane

Article abstract of DOI:10.1063/1.471214

The channels leading to the formation of atomic and/or molecular hydrogen in the photolysis of methane are studied. The H-atom elimination is investigated following photolysis at Lyman-α. The H₂ elimination is investigated following two-photon

Full text of DOI:10.1063/1.471214

Photofragment imaging of methane
Albert J. R. Heck, Richard N. Zare, and David W. Chandler
Citation: The Journal of Chemical Physics 104, 4019 (1996); doi: 10.1063/1.471214
View online: http://dx.doi.org/10.1063/1.471214
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/104/11?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Product state distributions in the dissociation of H3 (n=2,3) Rydberg states
J. Chem. Phys. 105, 3532 (1996); 10.1063/1.472220
Production of NH(ND) radicals in the reactions of N(22D) with H2(D2): Nascent vibrational distributions of NH(X
3Σ−) and ND(X 3Σ−)
J. Chem. Phys. 104, 9640 (1996); 10.1063/1.471716
Molecular dynamics in intense laser fields
AIP Conf. Proc. 295, 127 (1993); 10.1063/1.45233
Photodetachment measurements of the H− and D− density in a tandem multicusp ion source
AIP Conf. Proc. 287, 595 (1992); 10.1063/1.44835
Resonance enhanced twophoton dissociation of H2 by the X→(E F+G K+I) transition via intermediate B,C states
J. Chem. Phys. 95, 1608 (1991); 10.1063/1.461073
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.59.222.12 On: Mon, 01 Dec 2014 00:06:20

Photofragment imaging of methane
Albert J. R. Heck
Combustion Research Facility, Sandia National Laboratories, Livermore, California 94551
and Department of Chemistry, Stanford University, Stanford, California 94305
Richard N. Zare
Department of Chemistry, Stanford University, Stanford, California 94305
David W. Chandler
Combustion Research Facility, Sandia National Laboratories, Livermore. California 94551
͑
Received 6 October 1995; accepted 6 December 1995͒
The photolysis of methane is studied using photofragment imaging techniques. Our study reveals
that the photolysis of methane proceeds via many different pathways. The photofragment imaging
technique is used to resolve and characterize these various pathways and provides therefore unique
insight into the dynamical processes that govern this photodissociation. The formation of H-atom
photofragments following absorption of a Lyman-␣ photon, and H photofragments following
2
absorption of two ultraviolet photons ͑␭ϭ210–230 nm͒ are studied. The measured H-atom
photofragment images reveal that a channel that produces fast H atoms concomitant with methyl
fragments is dominant in the Lyman-␣ photolysis of methane. This channel leads to an anisotropic
recoil of the fragments. A secondary channel is observed leading to the formation of somewhat
slower H atoms, but an unique identification of this second channel is not possible from the data. At
least part of these slower H atoms are formed via a channel that produces H atoms concomitant with
CH and H photofragments. The recoil of these slower H atoms appears to be isotropic. The
2
measured, state-resolved H (v,J), photofragment images reveal that two channels lead to H
2
2
photofragments from the two-photon photolysis of methane: a channel that leads to H products
2
concomitant with methylene fragments; and a channel that leads to H products concomitant with
2
CH and H fragments. H (v,J) rotational and vibrational distributions are measured for each of these
2
two channels separately. The H products formed via the H ϩCH channel are rotationally and
2
2
2
vibrationally highly excited, whereas those formed via the H ϩCHϩH channel are rotationally and
2
vibrationally cooler. Rotational distributions of H formed via the H ϩCHϩH channel are well
2
2
reproduced by Boltzmann distributions. Results on D elimination following two-photon photolysis
2
of CD are in general similar and in qualitative agreement with the results on CH . © 1996
4
4
American Institute of Physics. ͓S0021-9606͑96͒03810-3͔
INTRODUCTION
methane are formed from 3s←1t excitations. The triply
2
1
degenerate T states corresponding to the 1t 3s electronic
2
2
Knowledge about the ͑photo͒ dissociation pathways of
the methane molecule is of fundamental importance as it is
of central importance to organic chemistry, and it is the pro-
totypical molecule having tetrahedral symmetry. Methane
photolysis is also of atmospheric interest as methane is quite
configuration are expected to be separated into three Jahn–
Teller components. Two components have been identified at
7,8
9
.7 and 10.4 eV. The excited states of methane and their
vertical excitation energies are not well characterized. It is
generally assumed that at photon energies corresponding to
abundant in the upper atmosphere of the earth and the atmo-
1
spheres of other planets.1–3 In the upper atmosphere of the
Lyman ␣ only the lowest singlet, T2, electronic state of
1
methane can be accessed. This T electronic state is be-
2
earth and in other planetary atmospheres, photolysis of meth-
ane is believed to be one of the primary steps in the synthesis
of higher hydrocarbons and other organic molecules. The
lowest lying electronic states of methane are located at high
energies above the ground state, corresponding in energy to
vacuum ultraviolet radiation. Therefore, the photochemistry
of methane in the atmosphere is mostly driven by intense
solar atomic emission lines, such as Lyman-␣ radiation. The
one-photon absorption spectrum of methane starts at ap-
proximately ␭Ͻ145 nm. The absorption cross section in-
creases continuously to shorter wavelengths, showing only
minor fine structure in the region between ␭ϭ100–145
lieved to have a short lifetime with respect to dissociation,
which partly explains the observed broad and diffuse absorp-
1
tion band. This T state correlates adiabatically with the
2
1
1
9,10
CH2 ͑ B ͒ϩH products.
Absorption in the region
1
2
␭ϭ105–115 nm ͑covered in this study͒ allows also excita-
tion to higher excited states. As mentioned before, a second
3
s←1t excitation is believed to have a threshold at ␭Ͻ119
2
nm, whereas a 3p←1t Rydberg excitation manifold is en-
2
7,11
ergetically allowed for ␭Ͻ112 nm.
Some theoretical data are available on excited states and
photolysis pathways of methane,9 but experimental data
are sparse. Very few experimental studies on methane pho-
tolysis have been performed under collision-free conditions.
,10
4
–6
nm.
This broad absorption region covers transitions to
several possible excited states. The lowest excited states of
J. Chem. Phys. 104 (11), 15 March 1996
0021-9606/96/104(11)/4019/12/$10.00
© 1996 American Institute of Physics
4019
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.59.222.12 On: Mon, 01 Dec 2014 00:06:20

4020
Heck, Zare, and Chandler: Photofragment imaging of methane
An important variable in atmospheric modeling of the role of
methane and its photoproducts is the quantum yields of
1
atomic and molecular hydrogen. Several channels in the
photolysis of methane may lead to H and H products. Con-
2
sidering only spin-conserving dissociation channels the fol-
lowing paths may be anticipated:
1
2
CH ͑ A ͒→CH ͑ AЉ͒ϩH ⌬Eϭ4.48 eV,
͑1͒
͑2͒
͑3͒
4
1
3
2
1
1
CH ͑ A ͒→CH ͑a A ͒ϩH ⌬Eϭ5.01 eV,
4
1
2
1
2
1
3
CH ͑ A ͒→CH ͑ B ͒ϩHϩH ⌬Eϭ9.21 eV,
4
1
2
1
1
1
CH ͑ A ͒→CH ͑a A ͒ϩHϩH ⌬Eϭ9.48 eV, ͑4͒
4
1
2
1
1
2
CH ͑ A ͒→CH͑ ⌸͒ϩH ϩH ⌬Eϭ9.07 eV.
͑5͒
4
1
2
Here, ⌬E is the energy required for each particular channel.
All energetic data used to obtain the reaction enthalpies are
taken from Lias et al.12 and the singlet–triplet splitting in
methylene is taken from McKellar et al.13 Earlier results on
the photolysis of methane under bulb conditions have been
reviewed by Ausloos and Lias.14 Quantum yields following
absorption at Lyman ␣ have been measured by Laufer and
FIG. 1. Schematic energy diagram showing the energetics of possible path-
ways in the photolysis of methane. The gray arrow indicates the excitation
energy deposited in the methane molecule following absorption of a
Lyman-␣ photon; the two black arrows indicate the excitation energy range
following absorption of two ultraviolet photons ͑␭ϭ210–230 nm͒.
15
16
McNesby for H , ␾ϭ0.58, Rebbert and Ausloos for CH,
2
17
␾
ϭ0.06, and Slanger and Black for H, ␾ϭ1.14. Addition-
ally, Slanger and Black concluded from their results that the
channel producing H atoms concomitant with H and CH,
2
channel ͑5͒, is active, and also that the contribution from
channel ͑1͒, leading to H atoms concomitant with methyl
fragments is only minor. Mordaunt et al.18 have studied the
formation of H-atom photofragments following absorption of
two-photon absorption ͑␭ϭ210–230 nm͒. In Fig. 1 an en-
ergy diagram is given which shows schematically the reac-
tion energetics for various channels expected in the photoly-
sis of methane for the photon energies used in our study. In
our study rotational and vibrational distributions of the H2
photofragments are measured, using resonance-enhanced
19
Lyman ␣ using H-atom Rydberg time-of-flight techniques.
1
8
Mordaunt et al. found a value of ␾ϭ1.0Ϯ0.5 for H. The
H-atom Rydberg time-of-flight technique also allowed them
to measure the kinetic energy release in the photofragments.
From these results it was concluded that a very significant
part of the H-atom photofragments in the photolysis of meth-
ane following absorption at Lyman ␣ is formed concomitant
with methyl radicals, in sharp contrast with results of Slanger
multiphoton ionization ͑REMPI͒ to detect the H (v,J)
2
photofragments. Additionally, angular and speed distribu-
tions of the H and H (v,J) photofragments are measured,
2
20
using two-dimensional photofragment imaging techniques.
The H and H (v,J) photofragment images allow us to deter-
2
mine and characterize the relative contributions of the chan-
nels active in the photolysis of methane. We also present
results on the D elimination following photolysis of CD .
1
7
and Black. Mordaunt et al. observed that some of the
H-atom photofragments are formed via alternative channels.
Although they could not directly assign which of the other
possible channels leads to these other H-atom photofrag-
ments, they suggested, using RRKM modeling, that decom-
position of metastable, internally excited methyl fragments,
contributes significantly. Collision-free studies on the forma-
tion of molecular hydrogen in the photolysis of methane are
even more sparse. Although it is generally assumed that
2
4
These latter results are similar and in qualitative agreement
with our results on CH4.
EXPERIMENT
The photofragment imaging apparatus used in this study
has been described elsewhere in detail.2
0,21
Neat methane or
perdeuteromethane ͑backing pressure 6 psig ͓110 kPa͔͒ is
expanded supersonically into a source vacuum chamber
through a solenoid valve ͑General Valve series 9͒. Approxi-
mately 1 cm from the nozzle orifice, the beam is skimmed
͑Beam Dynamics, orifice 0.8 mm͒ and collimated by a hole
͑diameter 1 mm͒ in the repeller plate.
H-atom elimination of methane is studied following ab-
sorption of a Lyman-␣ photon ͑121.6 nm͒. ͑1ϩ1Ј͒ Multipho-
ton ionization of H or D atomic photofragments is accom-
plished by absorption of a vacuum ultraviolet ͑vuv͒ Lyman-␣
and ultraviolet ͑uv͒ photon ͑364.8 nm͒. The vuv photon reso-
nantly excites the ground-state H atoms to the 2p state. The
more than one channel is active in the formation of H in the
2
photolysis of methane,1 little evidence exists to prove this
,14
assertion.
Measurements of total quantum yields of H and H in the
2
photolysis of methane are not adequate to unravel the dy-
namics that lead to these products as various different chan-
nels ͑1͒–͑5͒ are expected to contribute. In the present work
we investigate the channels that lead to formation of atomic
and/or molecular hydrogen in the photolysis of methane spe-
cifically. The H-atom elimination is studied following pho-
tolysis at Lyman ␣. The H elimination is studied following
2
J. Chem. Phys., Vol. 104, No. 11, 15 March 1996
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.59.222.12 On: Mon, 01 Dec 2014 00:06:20

Heck, Zare, and Chandler: Photofragment imaging of methane
4021
uv light is generated by frequency mixing the output of an
injection-seeded Nd:YAG-pumped dye laser ͑Spectra Phys-
ics GCR5, PDL 2͒ with the Nd:YAG infrared fundamental
light, using KD*P crystals. The bandwidth of the dye laser
Ϫ1
light is approximately 0.2 cm . The vuv Lyman-␣ light is
generated by frequency tripling of the uv light, focused ͑f
ϭ20 cm͒ into a gas cell, in a mixture of krypton and argon
adjusted to achieve optimum phase matching.2
1,22
The vuv
and residual uv laser light pass through a LiF lens ͑fϭ12.7
cm͒ into the vacuum chamber where they intersect the mo-
lecular beam.
H or D elimination is studied following absorption of
2
2
two uv photons ͑␭ϭ210–230 nm͒. The laser beam used to
dissociate methane by two-photon absorption is also used to
multiphoton ionize the H (v,J) or D (v,J) photofragments
2
2
state selectively. H and D are detected using ͑2ϩ1͒ REMPI
2
2
1
ϩ
23–26
spectroscopy via the E,F ⌺g state.
Ultraviolet light
with ␭Ͻ214 nm is generated by frequency doubling the
injection-seeded Nd:YAG-pumped dye laser light using
KD*P crystals. A DCM dye dissolved in methanol is used.
Subsequently, the frequency-doubled dye output is sum fre-
quency mixed with the residual dye laser output using appro-
priate BBO crystals. Ultraviolet light with ␭Ͼ214 nm is gen-
erated by first frequency doubling the injection-seeded
Nd:YAG-pumped dye laser using KD*P crystals. In this case
Fluorescein and LDS 590 dissolved in methanol are used.
Subsequently, the frequency-doubled dye output is sum fre-
quency mixed with the injection-seeded Nd:YAG fundamen-
tal infrared light ͑␭ϭ1.064 ␮m͒ using appropriate BBO crys-
tals. Throughout the experiments the average uv laser power
is kept at approximately 1 mJ per pulse. The uv light is
focused ͑fϭ15 cm͒ into the molecular beam.
In our experiments the direction of the polarization vec-
tor of the laser is chosen to be vertically polarized, i.e., par-
allel to the face of the position-sensitive ion detector. Fol-
lowing ionization, ions are accelerated along a 20 cm long
flight tube and impinge on a two-dimensional, position-
sensitive detector. The detector consists of a pair of chevron
microchannel plates ͑Galileo, 7.62 cm diameter͒ coupled to a
fast phosphor screen ͑P47 Phosphor, 50 ns decay time͒. The
laser, molecular beam, and detection system are all pulsed at
FIG. 2. H-atom photofragment ion image following absorption of a
Lyman-␣ photon in methane. This image was symmetrized ͑averaged over
its four redundant quadrants in order to improve the signal-to-noise ratio of
the image͒. This image is a two-dimensional projection of the recoiling H
photofragments.
RESULTS
H atom elimination following absorption at Lyman ␣
H-atom photofragments are observed when the com-
bined vuv ͑␭ϭ121.6 nm͒ and uv ͑␭ϭ365 nm͒ laser beams
intersect with the methane molecular beam. Figure 2 shows
the corresponding H-atom photofragment image. To record
the image, the laser is scanned over the Doppler profile dur-
ing data collection. To correct for the background H-atom
signal an image recorded with the laser firing well after the
molecular beam gas pulse is subtracted. The H-atom photo-
fragment image reveals that ͑1͒ two different channels con-
tribute and ͑2͒ the recoil of the photofragments is isotropic
for the inner channel, but anisotropic for the outer channel.
The recorded photofragment images are two-dimensional
projections of the three-dimensional recoil of the ionized
photofragments. The experimental configuration used dic-
tates that the initial three-dimensional distribution has cylin-
drical symmetry and therefore the inverse Abel transform27
can be used to reconstruct, from the two-dimensional image,
the intensity distribution through the center of the initial
three-dimensional ion cloud.20 The position of the impinging
ions on the position-sensitive detector can be converted to
velocity by measuring the displacement of the ions from the
center of the image and their arrival time. Velocity distribu-
tions are obtained from the inverse Abel transformed image
by scaling, appropriately, each pixel of the image.20
30 Hz. Timing is synchronized by digital delay generators
͑Stanford Research Systems͒. To record photofragment ion
images the front of the detector is pulsed to detect only one
mass peak. The resulting images are recorded using a Peltier-
cooled charge-coupled-device ͑CCD͒ camera ͑Photometrics,
576ϫ384 pixels͒ equipped with a standard 50 mm lens ͑Pen-
tax͒. Signal averaging is performed by opening the CCD
camera shutter, for approximately 12 000 laser shots, to
record a single image. The integrated charge on the CCD is
read by the associated camera electronics and transferred for
storage and processing to a MIPS RS 3000 computer. Back-
ground subtraction is achieved using an image recorded with
the laser firing well after the molecular beam gas pulse ͑typi-
cally 1 ms delay͒. Research grade methane was purchased
from Spectra Gasses ͑Vista, CA͒, perdeuteromethane ͑D4
The three-dimensional speed distribution of H-atom
photofragments from methane is shown in Fig. 3͑A͒ and re-
veals at least two distinct features. Fast H atoms are observed
with speeds that peak around 16 km/s and relatively slower
H atoms are observed with speeds that peak around 8 km/s.
Under the assumption that the H-atom photofragments are
formed concomitant with methyl fragments and using mo-
99%͒ from Cambridge Isotope laboratories ͑Andover, MA͒.
Both gases were used without further purification.
J. Chem. Phys., Vol. 104, No. 11, 15 March 1996
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.59.222.12 On: Mon, 01 Dec 2014 00:06:20

4022
Heck, Zare, and Chandler: Photofragment imaging of methane
distribution of the H-atom photofragments in the outer ring is
shown in Fig. 3͑C͒.
H elimination following two-photon absorption
2
In methane two-photon absorption, in the range ␭ϭ210–
2
30 nm, leads to H photofragments. Photofragment images
2
of H (v,J) were recorded state selectively for ͑vϭ0, Jϭ11,
2
1
3 and 15; vϭ1, Jϭ0–19; and vϭ2, Jϭ0–11͒. For CD4
photofragment images were recorded of D (v,J) state selec-
2
tively for ͑vϭ1, Jϭ16–19; and vϭ2, Jϭ0–10͒. Rotational
distributions are obtained by integrating the ion signal inten-
sities over the entire H /D (v,J) images. In order to obtain
2
2
rotational distributions corrections to these measured signal
intensities must be made which take into account nuclear
spin statistics and the differences in the transition moments
of the different ͑2ϩ1͒ REMPI transitions. Nuclear spin sta-
tistics lead to a 3:1 intensity alternations between odd and
even J levels of H and to a 1:2 intensity alternation between
2
odd and even J levels of D . The relative magnitudes of the
2
transition dipole moments in the ͑2ϩ1͒ REMPI scheme are
2
4
known from calibration experiments,
and ab initio
calculations.25 In general the calculated results are in good
agreement with the experimental data, but cover a wider
range of J values and therefore are used here to correct the
observed signal intensities. Measured H (v,J) signal inten-
2
sities, and corrected rotational distributions are given in
Table I. Measured D (v,J) signal intensities, and corrected
2
rotational distributions are given in Table II. In order to mea-
sure H (v,J) and D (v,J) photofragment images the laser
2
2
was scanned over the Doppler profile. In Fig. 4 photofrag-
ment images are shown for H ͑vϭ1͒, Jϭ1, 11, and 15. The
2
angular distributions of all measured H (v,J) products ap-
2
pear to be isotropic. The images shown in Fig. 4 indicate that
at least two distinct channels contribute. H speed distribu-
2
tions derived from the inverse Abel transformed images are
shown in Figs. 5 and 6, for vϭ1 and 2, respectively. The two
features observed in the images are seen in the speed distri-
butions as well. The channel that leads to relatively slow
H (v,J) products dominates the images at lower J levels,
2
whereas the second channel that leads to relatively faster
H (v,J) products increases in relative importance with in-
2
FIG. 3. ͑A͒ Speed distribution of H photofragments following absorption of
a Lyman-␣ photon in methane. ͑B͒ Kinetic energy release spectrum obtained
from the speed distributions of H-atom photofragments. The arrows indicate
the energetic thresholds for different channels ͑1͒, ͑3͒–͑5͒ in the Lyman-␣
photolysis of methane. ͑C͒ Angular distribution of the H-atom photofrag-
ments in the outer ring of the image shown in Fig. 2. This angular distribu-
tion was obtained by averaging over the H-atom photofragments with
speeds between 12 and 30 km/s.
creasing J. For H (vϭ1),Jу15, and for H (vϭ2),Jу10
2
2
only the channel leading to the relatively faster H products
2
is observed.
The formation and detection of H (v,J) photofragments
2
is the result of a multiphoton absorption process. To form H2
products methane molecules must absorb two photons. The
͑2ϩ1͒ REMPI detection scheme requires absorption of three
additional photons by the H products. The dependence of
2
the total H (v,J) ion signal on the laser fluence is shown in
2
mentum conservation arguments, the total kinetic energy re-
lease into the fragments can be derived from the H-atom
speed distributions. The kinetic energy release distribution
Fig. 7 for H ͑vϭ1, Jϭ9͒. Figure 7 reveals that the H ion
2
2
intensity is proportional to the laser fluence to the power
3.7Ϯ0.4. The ͑2ϩ1͒ REMPI detection has been found to be
dependent on the laser fluence to the power of 1.5Ϯ0.3.24
Therefore, a laser fluence dependence of approximately two
corresponding to the H-atom elimination from CH is shown
4
in Fig. 3͑B͒. In Fig. 3͑B͒ energetic thresholds are indicated
for several channels producing H atoms from the photolysis
of methane following absorption of Lyman ␣. The angular
is observed for the H product formation, as expected for a
2
nonresonant two-photon absorption process. Photofragment
J. Chem. Phys., Vol. 104, No. 11, 15 March 1996
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.59.222.12 On: Mon, 01 Dec 2014 00:06:20

Heck, Zare, and Chandler: Photofragment imaging of methane
TABLE I. Rotational distributions H (v,J) photofragments from CH .
4023
2
4
EvibϩErot
internal energy intensities Rotational total intensities Fraction in distribution distribution
Measured
Corrected
Fast channel Slow channel
J
͑eV͒
͑arb. units͒ correction
͑arb. units͒
fast channel ͑arb. units͒
͑arb. units͒
H ͑vϭ0͒
2
1
1
0.872
1.15
1.44
1.72
2.26
2.58
2.43
1.76
0.911
0.867
0.812
0.812a
2.06
2.24
1.98
1.43
0.39
0.55
0.69
0.92
0.793
1.23
1.37
1.31
1.27
1.01
0.608
0.114
1
1
1
3
5
7
H ͑vϭ1͒
2
0
1
2
3
4
5
6
7
8
9
0.516
0.530
0.558
0.599
0.654
0.721
0.800
0.890
0.991
1.10
1.22
1.34
1.47
1.61
1.74
1.88
2.01
2.14
0.730
5.85
2.44
8.45
3.39
1.09
1.08
1.08
1.07
1.05
1.04
1.02
1.01
0.987
0.967
0.945
0.921
0.959
0.870
1.09
0.806
0.806a
0.806a
0.806a
0.806a
2.39
6.32
7.91
9.04
10.7
10.4
11.1
10.1
10.4
9.14
8.51
9.18
8.40
6.56
3.47
3.99
2.78
2.56
1.86
1.03
0.26
0.26
0.27
0.29
0.31
0.33
0.39
0.43
0.47
0.48
0.57
0.66
0.75
0.89
0.611
1.63
2.17
2.67
3.37
3.44
4.30
4.32
4.88
4.41
4.84
6.08
6.27
5.83
1.78
4.69
5.74
6.37
7.33
6.96
6.80
5.75
5.52
4.73
3.67
3.11
2.13
0.731
10.0
3.63
9.97
3.52
9.45
3.00
9.97
2.92
7.54
1.06
4.95
1.15
3.17
0.770
1.28
10
1
1
12
13
14
15
16
17
18
19
1.0
1.0
1.0
1.0
1.0
3.99
2.78
2.56
1.86
1.03
0.00
0.00
0.00
0.00
0.00
2.27
2.39
H ͑vϭ2͒
2
0
1
2
3
4
5
6
7
8
9
1.00
1.02
1.04
1.08
1.13
1.20
1.27
1.36
1.45
1.56
1.67
1.79
0.950
5.53
2.37
8.81
3.12
10.1
3.16
10.4
3.20
8.18
1.98
5.13
1.06
1.05
1.04
1.04
1.02
1.01
0.989
0.971
0.946
0.921
0.897
0.869
3.01
5.81
7.43
9.13
9.58
10.2
9.38
10.1
9.08
7.53
5.33
4.46
0.53
0.56
0.56
0.58
0.62
0.68
0.71
0.78
0.85
0.92
1.0
1.60
3.23
4.18
5.33
5.92
6.92
6.63
7.91
7.76
6.93
5.33
4.46
1.41
2.58
3.25
3.80
3.66
3.26
2.75
2.19
1.32
0.603
0.00
0.00
10
11
1.0
aRotational correction factors are unknown. As an approximate value the rotational correction factor of the
nearest J level is used.
images were recorded at different laser fluences for H ͑vϭ1,
ages reveal directly that there are two distinct channels. The
speed distributions obtained in our experiments, which are
shown in Fig. 3͑A͒, clearly show the two channels in the
formation of H-atom photofragments. Following absorption
of a ␭ϭ121.6 nm photon several dissociation channels are
energetically accessible. The least endothermic channel ͑1͒,
which produces H atoms concomitant with methyl radicals in
2
Jϭ9͒. The H speed distributions derived from the images,
2
taken at laser fluences of 500, 700, and 1000 ␮J, are shown
in Fig. 8. Within the error of the measurements these speed
distributions are identical. We conclude that the two ob-
served channels have the same dependence on laser fluence.
DISCUSSION
2
the AЉ electronic ground state, is endothermic by 4.48 eV.
2
H atom elimination following absorption at Lyman ␣
Therefore, following absorption of a Lyman-␣ photon the
Mordaunt et al.18 have recently studied the H-atom
maximum possible amount of kinetic energy released into
the H and CH fragments is approximately 5.7 eV. The fast-
elimination in the Lyman-␣ photodissociation of methane us-
ing the H-atom Rydberg time-of-flight technique. The
present study on the H-atom elimination from methane has a
similar objective. However, we use photofragment imaging
to obtain in a single experiment the speed and angular dis-
tributions of the H-atom photofragments. Our H-atom im-
3
est H atoms observed in the H-atom photofragment images
have speeds that corresponds to total fragment kinetic energy
releases of approximately 5 eV ͑see Fig. 3͒. However, the
average speed of the H atoms in the outer channel corre-
sponds to a total fragment kinetic energy of approximately
J. Chem. Phys., Vol. 104, No. 11, 15 March 1996
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.59.222.12 On: Mon, 01 Dec 2014 00:06:20

4024
Heck, Zare, and Chandler: Photofragment imaging of methane
TABLE II. Rotational distributions D (v,J) photofragments from CD .
2
4
EvibϩErot
internal energy intensities Rotational total intensities Fraction in distribution distribution
Measured
Corrected
Fast channel Slow channel
J
͑eV͒
͑arb. units͒ correction
͑arb. units͒
fast channel ͑arb. units͒
͑arb. units͒
D ͑vϭ1͒
2
16
17
18
19
1.25
1.34
1.44
1.54
3.240 0
1.670 0
3.370 0
1.440 0
0.880
0.870
0.870a
0.870a
2.85
2.90
2.93
2.51
0.60
0.62
0.72
0.74
1.72
1.79
2.11
1.86
1.13
1.11
0.820
0.644
D ͑vϭ2͒
2
0
1
2
3
4
5
6
7
8
9
0.728
0.734
0.748
0.769
0.796
0.830
0.870
0.917
0.969
1.03
0.845 00
0.972 00
2.636 8
1.952 0
5.038 0
2.754 0
5.574 1
2.639 0
5.272 8
2.701 0
5.346 0
1.03
1.03
1.02
1.02
1.01
1.01
0.998
0.988
0.977
0.965
0.952
0.870
2.00
2.70
3.98
5.10
5.54
5.56
5.21
5.15
5.21
5.09
0.34
0.34
0.34
0.36
0.38
0.39
0.40
0.43
0.47
0.50
0.55
0.293
0.678
0.918
1.41
1.93
2.16
2.23
2.24
2.41
2.61
2.78
0.577
1.32
1.78
2.57
3.17
3.38
3.33
2.97
2.74
2.61
2.31
1
0
1.09
aRotational correction factors are unknown. As an approximate value the rotational correction factor of the
nearest J level is used.
1.5 eV. All the H atoms in the outer, fast channel are formed
results for the H elimination, which are discussed below, do
2
exclusively concomitant with electronic ground state CH3
indicate that channel ͑5͒ is a significant channel in the pho-
2
(
AЉ) fragments: Based on energy conservation other chan-
tolysis of methane. We estimate that the slow channel con-
tributes approximately 13%Ϯ5% to the total of the H-atom
photofragments in the photolysis of methane at Lyman ␣.
This estimate implies that the rest, 87%Ϯ5% of the H atoms
are formed concomitant with methyl radicals.
The images show that the faster H-atom photofragments
result from an anisotropic recoil, whereas the recoil of the
slower H-atom photofragments appears to be isotropic. The
angular distribution of photofragments can be described by28
2
nels and/or the formation of electronically excited methyl
fragments cannot lead to these fast H atoms. The deduced,
average amount of internal energy deposited in the CH frag-
3
ments, as derived from the kinetic energy release spectrum,
is high, i.e., approximately 4.2 eV. Thus for the outer chan-
nel, the fraction of available energy deposited in translational
energy, ftrans , is 0.25, whereas the fraction deposited in in-
ternal energy, f ϭfvibϩfrot , is 0.75.
int
Methyl fragments with more than 4.7 eV of internal en-
I͑␪͒Ϸ1ϩ␤ P ͑cos ␪͒,
͑6͒
2
2
ergy may decompose spontaneously into CHϩH or when a
2
little more internal energy is available into CH ϩH. Interest-
where P ͑cos ␪͒ is the second order Legendre polynomial
2
2
ingly, the outer channel in the speed distribution drops
sharply at the threshold energies for these processes, i.e., for
kinetic energy releases below 1 eV. This drop may be an
indication that one or more of the channels ͑3͒, ͑4͒, and ͑5͒
start to contribute to the formation of H-atom photofrag-
ments at their energetic thresholds ͓see Fig. 3͑B͔͒. Further
decomposition of highly excited methyl fragments may lead
and ␤ is the anisotropy parameter. In Fig. 3͑C͒ the experi-
2
mentally measured angular distribution of the H-atom
photofragments in the outer ring of the image in Fig. 2, is
shown together with a fit to this data using Eq. ͑6͒. A best fit
to the angular distribution of the H-atom photofragments in
the outer ring of the image is obtained with an anisotropy
parameter, ␤ , of 0.28. The observed anisotropy of the fast
2
to additional H and/or H photofragments via channels ͑3͒,
channel may be caused by the symmetry of the electronic
states involved and/or by the relatively short lifetime of the
excited methane molecules, i.e., when the methane mol-
ecules rotate substantially prior to dissociation anisotropy of
the photofragments will be partly washed out.
2
͑4͒, and ͑5͒. These H atoms can be formed, in principle, in a
simultaneous or stepwise breaking of bonds in methane. If
the secondary H atoms are formed via a stepwise process the
calculated fragment kinetic energy release in Fig. 3͑B͒ must
be slightly adjusted, but by no more than 3%. The H-atom
photofragment image and speed distribution clearly reveal a
second channel that produces relatively slower H atoms. As
shown in Fig. 3͑B͒ the onset of this channel corresponds
closely with the thresholds of channels ͑3͒, ͑4͒, and ͑5͒.
These three channels, which all produce H atoms, have
nearly the same energetic thresholds. From our H-atom
photofragment images we are not able to resolve which of
these channels contributes to the observed signal, but our
Our results and those of Mordaunt et al.18 both reveal
two distinct channels in the H-atom elimination from meth-
ane following absorption at Lyman ␣. We have measured the
speed and angular distributions of the photofragments and
find that the recoil of the fast H atoms is anisotropic, the
channel leading to slower H atoms is isotropic. By measur-
ing the complete H-atom photofragment three-dimensional
distribution we not only obtain the angular and speed distri-
butions of the H-atom photofragments but we also are able to
J. Chem. Phys., Vol. 104, No. 11, 15 March 1996
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.59.222.12 On: Mon, 01 Dec 2014 00:06:20

Heck, Zare, and Chandler: Photofragment imaging of methane
4025
FIG. 5. H ͑vϭ1, J͒ speed distributions. Each frame shows the speed distri-
2
bution of a particular rotational level J. The gray lines are the observed
speed distributions, the dashed lines are Gaussian fit functions used to model
the speed distributions, and the solid black line is the resulting fit to the
speed distribution ͑see the text for details͒.
two-photon absorption and formation of singlet CH ϩH this
2
2
channel will lead to highly internally or translationally ex-
cited photofragments. The observed fast channel in the
H (v,J) photofragment images peaks at speeds of typically
2
1
3 km/s. This speed corresponds to a kinetic energy release
FIG. 4. H photofragment ion images following two-photon ͑uv͒ absorption
2
into the fragments of approximately 2 eV, substantially less
than the energetic limit of 5.7 eV. Based on energy conser-
vation the fast channel observed in the images can coincide
only with formation of CH ϩH . Our data reveal that chan-
͑␭ϭ210–230 nm͒ in methane. These images are two dimensional projec-
tions of the recoiling H2 photofragments. From the top to the bottom are
displayed the ion images of H ͑vϭ1, Jϭ1͒, H ͑vϭ1, Jϭ11͒, and H ͑vϭ1,
2
2
2
Jϭ15͒. On the right are shown intensity distributions along a vertical slice
through the center of the images.
2
2
nel ͑2͒ leads to high internal excitation of not only the H2
products, but also the CH fragments. The slower channel in
2
the H photofragment images corresponds with an average
kinetic energy release into the fragments of approximately
2
quantify the ratio of H atoms formed concomitant with me-
thyl radicals ͓channel ͑1͔͒ and H atoms formed via three-
body dissociation channels ͑3͒–͑5͒. We find that channel ͑1͒
is the most dominant channel leading to H-atom photofrag-
0
.25 eV. One of the most plausible explanations for this
channel is the formation of CHϩHϩH . This channel is
2
highly endothermic, but the large amount of photolysis en-
ergy still makes it accessible. Assuming that only two major
ments. Previous models used to model the chemistry of the
upper atmosphere and various planets1,29 have assumed that
channels contribute to the H (v,J) photofragment images the
2
channel ͑1͒ can be neglected in the photolysis of methane.
Our studies reveal that this assumption is not correct.
speed distributions are fitted to two Gaussian functions. As
variables the widths, heights and positions of the two Gaus-
sians were optimized in the fit. Some typical results obtained
H elimination following two-photon absorption
2
from these fits are shown in Figs. 5 and 6 for H , vϭ1 and
2
Two-photon absorption, which deposits between 10.8
and 11.8 eV of energy in the methane molecule, leads to H2
vϭ2, respectively. Using this procedure all H (v,J) and
2
D (v,J) speed distributions could be fitted satisfactorily. Fit-
2
photofragments. The recorded H (v,J) photofragment im-
ting the speed distributions allows us to determine the con-
tributions of the two different channels to the images. Tables
I and II show the derived fractions of products in the fast
channel for each measured H (v,J) and D (v,J) level. By
2
ages reveal that there are two distinct speed groups of H2
products. Ignoring the spin-forbidden formation of triplet
methylene, the least endothermic dissociation channel of
2
2
methane that leads to H products is channel ͑2͒. As approxi-
knowing the relative contributions of each of the two chan-
2
mately 5.7 eV is available for product excitation following
nels, the rotational distributions for each individual channel
J. Chem. Phys., Vol. 104, No. 11, 15 March 1996
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.59.222.12 On: Mon, 01 Dec 2014 00:06:20

4026
Heck, Zare, and Chandler: Photofragment imaging of methane
FIG. 8. H ͑vϭ1,9͒ speed distributions obtained from ion images measured
2
at three different laser powers of 500 ͑open circles͒, 700 ͑filled circles͒, and
1000 ␮J ͑gray triangles͒, respectively.
11.5 eV. As mentioned in the Introduction this amount of
excitation energy allows excitation to different excited states
of methane.5 We do not know exactly which states are
excited in the two-photon process but from the smooth rota-
,7,8
FIG. 6. H ͑vϭ2, J͒ speed distributions. Each frame shows the speed distri-
2
bution of a particular rotational level J. The solid lines are the observed
speed distributions, the dashed lines are Gaussian fit functions used to model
the speed distributions, and the solid black line is the resulting fit to the
speed distribution ͑see the text for details͒.
also can be obtained. These rotational distributions are
shown in Fig. 9 for H ͑vϭ1 and 2͒ and in Fig. 10 for
2
D ͑vϭ2͒.
2
The photolysis laser is tuned over an energy range of
approximately 1 eV to state-selectively detect the various
H (v,J) products. The photolysis energy ranges from 10.5 to
2
FIG. 9. H2 rotational distributions in the photolysis of methane for vϭ1
͑
top͒ and vϭ2 ͑bottom͒. The rotational distributions obtained by integrating
the ion signals over the whole of the images are displayed by open circles.
The rotational distributions of the fast and slow H channels in the photoly-
sis of methane are shown by filled circles and filled triangles, respectively.
2
ϩ
FIG. 7. Log/log plot of the H2 signal intensity versus the laser fluence.
J. Chem. Phys., Vol. 104, No. 11, 15 March 1996
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.59.222.12 On: Mon, 01 Dec 2014 00:06:20

Heck, Zare, and Chandler: Photofragment imaging of methane
4027
FIG. 10. D rotational distributions in the photolysis of perdeuteromethane
2
for vϭ2. The rotational distributions obtained by integrating ion signals
over the whole of the images are displayed by open circles. The rotational
distributions of the fast and slow D channels in the photolysis of CD are
2
4
shown by filled circles and filled triangles, respectively.
tional distributions observed we assume that the excitation is
fairly broad and unstructured. The one-photon absorption
spectrum in this region is fairly continuous, but displays
some minor structure around 11 eV, attributed to vibrational
5
levels of the 3p Rydberg state. The two-photon absorption
spectrum in this energy region is unknown. Galasso30 has
calculated transition probabilities for multiphoton absorption
processes in methane. His calculations predict that in the
two-photon absorption process 3p←1t excitations are rela-
2
tively stronger than 3s←1t excitations. The observed
2
FIG. 11. ͑A͒ Percentages of the contribution of the fast channel to the
H (v,J) photofragment images in the two-photon photolysis of CH plotted
H (v,J) rotational distributions are fairly continuous, for
2
2
4
both channels. We interpret the smoothness of the distribu-
tions as an indication that the variation in photolysis energy
does not effect the observed dynamics significantly. Al-
versus the internal energy of the corresponding H (v,J) products ͑B͒ As in
2
͑
A͒, but now for D (v,J) products in the two-photon photolysis of CD .
2
4
though the H (v,J) products we have detected are formed
2
following absorption at different excitation energies, we be-
lieve they are produced via similar dynamic processes, i.e.,
via the same potential energy surfaces. A possible pathway
might be that initial excitation to specific excited states of
methane is followed by fast internal conversion to the S0
ground electronic state.
hydrogen concomitant with methylene, is less endothermic,
by approximately 4 eV, than the slower channel, which is
attributed to the formation of molecular hydrogen concomi-
tant with CH and H. This observed behavior gives additional
confidence to the assignments of the fast channel ͑2͒ and
slow channel ͑5͒. The linear correlation between the branch-
ing ratio and the H2 photofragment internal energy might
indicate that the two channels are coupled, i.e., the dynamics
that lead to channels ͑2͒ and ͑5͒ occur via the same potential
energy surfaces.
We derive the relative contributions for each of the two
channels from the H (v,J) and D (v,J) photofragment im-
2
2
ages and speed distributions. For all studied vibrational lev-
els of H and D the slower channel contributes relatively
2
2
more in the lower J levels than in the higher J levels. In Fig.
1 the contributions of the fast channel ͑in %͒ to the images
are plotted versus the internal energy ͑EvibϩE ͒ of the H2
1
ROTATIONAL DISTRIBUTIONS
The fast channel
rot
͓
Fig. 11͑A͔͒ and D ͓Fig. 11͑B͔͒ products. As seen in Fig. 11
2
an almost linear correlation exits, up to the cutoff of the
slower channel, between the internal energy of the hydrogen
molecules and the branching ratio of the two channels. This
linear correlation is approximately identical for the different
vibrational levels. The fact that the relative contribution of
the fast channel increases with increasing internal energy of
the products is somewhat counterintuitive. However, the fast
channel, which is attributed to the formation of molecular
Dictated by conservation of energy, the outer channel in
the images, corresponding to very fast H photofragments,
2
can only be attributed to the formation of molecular hydro-
gen concomitant with methylene. The rotational distributions
observed for this channel, displayed in Figs. 9 and 10, reveal
that the H photofragments are formed with high amounts of
2
rotational energy. For instance, the peak in the rotational dis-
tribution for this channel for H ͑vϭ1͒ is at Jϭ12, which
2
J. Chem. Phys., Vol. 104, No. 11, 15 March 1996
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.59.222.12 On: Mon, 01 Dec 2014 00:06:20

4028
Heck, Zare, and Chandler: Photofragment imaging of methane
corresponds to approximately 1 eV of rotational energy. Ap-
proximately 6 eV is available following two-photon absorp-
tion and the formation of ground-state CH ϩH . As men-
2
2
tioned above the average fragment kinetic energy release for
this channel is approximately 2 eV. Conservation of energy
implies that the methylene products formed via channel ͑2͒
are formed with approximately 3 eV of internal energy. The
methylene products may be formed in electronically excited
1
1
singlet states. The b B and the c A states of methylene
1
1
1
are approximately 0.9 and 3.4 eV above the a A state.
1
1
1
Emission from the b B to the a A state of methylene has
1
1
been observed following photolysis of methane in the
ϭ50–132 nm range.5 The fluorescence quantum yield of
,6
␭
this emission reaches its maximum at approximately ␭ϭ106
nm; however, even at this maximum the fluorescence quan-
6
tum yield is less than 1%. This is taken as evidence that the
methylene molecules are predominantly formed in the lowest
singlet state, but with large amounts of vibrational and rota-
tional energy. Some theoretical studies have appeared on the
reverse reaction, namely, the insertion of singlet methylene
3
1,32
into H2.
Singlet methylene reacts with molecular hydro-
33
gen without a significant activation barrier. When the tra-
jectories calculated to be the minimum energy path for this
reaction are run in the reverse order several interesting fea-
tures are observed. These trajectories predict that the disso-
ciation of methane into singlet methylene and molecular hy-
drogen follows a stepwise path.32 Following excitation a
hydrogen atom flies from the methane molecule and takes
with it a second hydrogen atom, to form a H fragment, as it
2
leaves. Such a path would lead to high rotational excitation
of the H products, because of the initial asymmetry of the
2
path. This mechanism is in agreement with the observed high
rotational excitation of the molecular hydrogen products.
The slow channel
We attribute the inner channel of the H images to the
2
channel that forms H products concomitant with CH and H
2
products. The H photofragments in this channel are formed
2
with relatively less rotational energy than those in the outer
channel, and the fragments’ kinetic energy releases do not
exceed 0.5 eV. No rotational population is observed in this
FIG. 12. Rotational distributions in the photolysis of CH /CD for the slow
channel at J levels above 14 and 9, for H ͑vϭ1͒ and
4
4
2
H /D product channel. From top to bottom are shown H ͑vϭ1͒, H ͑vϭ2͒,
2
2
2
2
H ͑vϭ2͒, respectively. Formation of H ͑vϭ1͒ via channel
2
2
and D ͑vϭ2͒ experimental rotational distributions ͑circles͒ and Boltzmann
2
͑5͒, i.e., concomitant with CH and H fragments, is energeti-
distributions with rotational temperatures ͑Trot͒ of 3850, 2150, and 2700 K,
cally not allowed for rotational levels with JϾ15, whereas
respectively.
formation of H ͑vϭ2͒ is energetically not allowed for JϾ10.
2
The observed cutoff in the rotational distributions are in
agreement with the assignment of the inner channel to chan-
nel ͑5͒. Figure 12 displays the rotational distributions of the
slow channel and Boltzmann distributions with rotational
ments. Mordaunt et al.18 have suggested the possibility that
in the photolysis of methane excited methyl radicals are
formed that may undergo further unimolecular decomposi-
temperatures, Trot , of 3850, 2150 and 2700 K for H ͑vϭ1͒,
2
H ͑vϭ2͒, and D ͑vϭ2͒, respectively. As can be seen in Fig.
tion into 3either CH ϩH or CHϩH2.18 Theoretical
2
2
2
4
35,36
12 these Boltzmann distributions reproduce quite well the
calculations as well as experimental data
indicate that
observed rotational distributions of the slow channel.
The mechanism for the slow channel can not uniquely be
determined from our data as we cannot distinguish between a
the minimum energy paths of both these decomposition re-
actions involves negligible activation energy. RRKM calcu-
lations performed by Mordaunt et al. indicate that the de-
composition of internally excited ground state methyl
stepwise or concerted formation of the HϩH ϩCH frag-
2
J. Chem. Phys., Vol. 104, No. 11, 15 March 1996
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.59.222.12 On: Mon, 01 Dec 2014 00:06:20

Heck, Zare, and Chandler: Photofragment imaging of methane
TABLE III. Vibrational populations ͑arb. units͒ H (v) photofragments in
4029
2
the photolysis of CH4 .
v
Fast channel
Slow channel
1
2
67.0
66.2
65.3
24.8
radicals into CHϩH is favored at least by almost 2 orders of
2
magnitude over the decomposition into CH ϩH over the in-
2
ternal energy range between 4.7 and 5.7 eV. From our data it
seems plausible that the H products observed in the slow
2
channel in the images and speed distributions are formed via
such a mechanism.
Summarizing, we find that the rotational distributions of
the fast and slow channel are markedly different. The fast
FIG. 13. H2 vibrational populations in the two-photon photolysis of CH4 .
The vibrational populations in the fast and slow H channels in the photoly-
2
sis of CH are shown by striped and filled bars, respectively.
4
channel in the formation of H products from the photolysis
2
of methane is attributed to the direct formation of molecular
hydrogen and singlet methylene. These H2 products are
formed with a very high amount of rotational energy. The
The fast channel, which leads to H products concomitant
with methylene, leads to a significant amount of vibrational
2
H (v,J) rotational distributions are smoothly shaped, and no
2
excitation in the H . From the data shown in Table III and
2
sharp cutoff or structure is observed. The concomitant meth-
ylene products are formed with large amounts of internal
Fig. 13 we infer that there will be population in vϾ2 levels
as well. All rotational levels that are energetically accessible
for the slow channel in both vϭ1 and 2 are probed. The slow
channel leads to much less vibrational excitation of the H2
products than the fast channel.
The relative importance of channels ͑2͒ and ͑5͒ in the
photolysis of methane can be deduced from our data for vϭ1
and vϭ2. It is observed that both channels ͑2͒ and ͑5͒ con-
excitation as well. Based on previous CH fluorescence
2
5
,6
studies we believe that most of the methylene products are
formed in their lowest singlet electronic state, but with high
amounts of vibrational and rotational excitation. The experi-
mentally observed rotational distributions in the slow chan-
nel are markedly cooler and can be modeled well with rota-
tional temperatures. Rotational population in the slow
channel is observed to vanish just below the energetic thresh-
tribute approximately equally to the formation of H prod-
2
ucts in vϭ1, whereas in vϭ2 channel ͑2͒ contributes ap-
proximately three times more than channel ͑5͒. We have not
probed all energetically accessible and/or populated rota-
tional and vibrational levels, and therefore we cannot deter-
mine the total yields of the two channels. Our data show,
however, that the importance of channel ͑2͒, i.e., the forma-
old for the formation of CHϩHϩH (v,J). The slow channel
2
is attributed to the formation of H concomitant with CH and
2
H.
VIBRATIONAL DISTRIBUTIONS
tion of methylene concomitant with H is of the same order
2
Our set of data is not sufficient to obtain full vibrational
as that of channel ͑5͒, i.e., the formation of H concomitant
2
distributions of the H photofragments in the photolysis of
2
with CHϩH.
methane. However, relative vibrational populations can be
obtained for H ͑vϭ1͒ and ͑vϭ2͒. These vibrational popula-
2
tions are obtained by summing the populations in all mea-
sured rotational levels for the slow and fast channel sepa-
rately. However, the vibrational populations are given in
Table III and shown graphically in Fig. 13. In general the
CONCLUSIONS
We have studied the photolysis of methane using photo-
fragment imaging to learn about the formation of H-atom
photofragments following absorption of a Lyman-␣ photon,
and the formation of H2 photofragments following two-
photon absorption ͑␭ϭ210–230 nm͒. We have determined
sensitivity for detection of H using the ͑2ϩ1͒ REMPI de-
2
tection scheme is different for each vibrational level. Typi-
cally, vibrational correction factors, which have been deter-
mined experimentally and theoretically,2
4,25
must be used to
the speed and angular distributions of the H and H photo-
2
obtain vibrational distributions. The vibrational correction
fragments, as well as rotational and vibrational distributions
factors are approximately equal for H , vϭ1 and 2, and
of the H photofragments. Our study reveals that the vacuum
2
2
therefore we apply no vibrational correction. For H ͑vϭ2͒
ultraviolet photolysis of methane proceeds via several differ-
ent pathways. The photofragment imaging technique allows
us to unravel and characterize various pathways that lead to
2
we expect some rotational population in J levels higher than
we have probed. The vibrational population in the fast chan-
nel is therefore somewhat underestimated for vϭ2.
H and H photofragments and provides new insight into the
2
For the fast channel the H vibrational population is ap-
proximately equal in vϭ1 and 2. For the slow channel the
vibrational population in vϭ1 is much higher than in vϭ2.
dynamical processes that govern this photodissociation.
We find that the most significant channel in the forma-
tion of H-atom photofragments in the photolysis of methane
2
J. Chem. Phys., Vol. 104, No. 11, 15 March 1996
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.59.222.12 On: Mon, 01 Dec 2014 00:06:20

4030
Heck, Zare, and Chandler: Photofragment imaging of methane
is channel ͑1͒, i.e., the formation of H atoms concomitant
with methyl fragments. The methyl fragments formed via
this channel are hot and have an average internal energy of
approximately 4 eV. The recoil of the photofragments formed
via this channel is anisotropic and can be described by an
anisotropy parameter, ␤, of 0.28. A second channel is ob-
served to lead to slower H-atom photofragments. These
H-atom photofragments are formed via the channels ͑3͒, ͑4͒,
and/or ͑5͒. The recoil of these slower H-atom photofrag-
ments is isotropic. From our H-atom photofragment images
alone, we cannot directly unravel to which extent channels
ACKNOWLEDGMENTS
A.J.R.H. thanks the Associated Western Universities,
Inc. for an AWU-DOE fellowship. This work was supported
by the U.S. Department of Energy, Office of Basic Energy
Sciences, Division of Chemical Sciences and in part by the
U.S. National Science Foundation under NSF CHE-9322690.
1
J. S. Levine, The Photochemistry of Atmospheres ͑Academic, New York,
1985͒.
2
R. P. Wayne, Chemistry of Atmospheres ͑Oxford University, New York,
991͒.
1
3
4
5
6
7
L. M. Lara, R. D. Lorenz, and R. Rodrigo, Planet. Space Sci. 42, 5 ͑1994͒.
H. Okabe, Photochemistry of Small Molecules ͑Wiley, New York, 1978͒.
L. C. Lee and C. C. Chiang, J. Chem. Phys. 78, 688 ͑1983͒.
G. Ma, M. Suto, and L. C. Lee, J. Quant. Radiat. Transfer. 44, 379 ͑1990͒.
M. B. Robin, Higher Excited States of Polyatomic Molecules ͑Academic,
New York, 1974͒.
͑3͒, ͑4͒, and/or ͑5͒ contribute.
The H (v,J) photofragment images reveal that there are
2
two channels leading to molecular hydrogen in the photoly-
sis of methane. From the images we determine the relative
contributions from these channels for each H (v,J) product.
8
2
M. G. Curtis and I. C. Walker, J. Chem. Soc., Fraday Trans. 2. 85, 659
͑1989͒.
One channel leads to relatively fast H products, which are
2
9
0
formed simultaneous with CH fragments. The rotational dis-
S. Karplus and R. Bersohn, J. Chem. Phys. 51, 2040 ͑1969͒.
M. S. Gordon and J. W. Caldwell, J. Chem. Phys. 70, 5503 ͑1979͒.
H. U. Lee and R. Janoschek, Chem. Phys. 39, 271 ͑1979͒.
S. G. Lias, J. E. Bartmess, J. F. Liebman, J. L. Holmes, R. D. Levin, and
W. G. Mallard, J. Phys. Chem. Ref. Data. 17, suppl. 1, ͑1988͒.
A. R. W. McKellar, P. R. Bunker, T. J. Sears, K. M. Evenson, R. J.
Saykally, and S. R. Langhoff, J. Chem. Phys. 79, 5151 ͑1983͒.
P. J. Ausloos and S. G. Lias, Annu. Rev. Phys. Chem. 22, 85 ͑1971͒.
A. H. Laufer and J. R. McNesby, J. Chem. Phys. 49, 2272 ͑1968͒.
R. E. Rebbert and P. Ausloos, J. Photochem. 1, 171 ͑1972͒.
T. G. Slanger and G. Black, J. Chem. Phys. 77, 2432 ͑1982͒.
D. H. Mordaunt, I. R. Lambert, G. P. Morley, M. N. R. Ashfold, R. N.
Dixon, L. Schineider, and K. H. Welge, J. Chem. Phys. 98, 2054 ͑1993͒.
M. N. R. Ashfold, I. R. Lambert, D. H. Mordaunt, G. P. Morley, and C. M.
Western, J. Phys. Chem. 96, 2938 ͑1992͒.
2
1
tributions of the H (v,J) products in this channel ͑2͒ are
2
11
12
very hot, i.e., H (v,J) products with more than 1.8 eV of
2
rotational energy are observed. The methylene products
formed via channel ͑2͒ are formed with an average of ap-
proximately 3 eV of internal energy. The second channel
1
1
3
4
leading to H products is most probably channel ͑5͒, i.e., the
15
2
1
1
1
6
7
8
formation of H products concomitant with CH and H. This
2
assignment is consistent with the observed cutoff in the ro-
tational distributions. The rotational distributions for this
channel are markedly cooler and can be described by Boltz-
19
mann rotational temperatures of T ϭ3850 and 2150 K for
rot
20
A. J. R. Heck and D. W. Chandler, Annu. Rev. Phys. Chem. 46, 335
vϭ1 and 2, respectively. The observation of this channel in
͑
1995͒.
the formation of H products implies that this channel is also
2
21
T. N. Kitsopoulos, M. A. Buntine, D. P. Baldwin, R. N. Zare, and D. W.
Chandler, Science. 260, 1605 ͑1993͒.
R. Hilbig and R. Wallenstein, IEEE J. Quantum Electron. QE-19, 194
responsible for the formation of H-atom photofragments. The
2
2
2
3
ratio of these two channels, which both lead to H products is
2
͑
1983͒.
E. E. Marinero, R. Vasudev, and R. N. Zare, J. Chem. Phys. 78, 692
1983͒.
found to be strongly dependent on the internal energy of the
H (v,J) products.
2
͑
In summary, our results on the photolysis of methane
reveal that channel ͑1͒, i.e., the formation of H atoms con-
comitant with methyl fragments, is far more important, by a
factor of 6, than the summed contributions of all alternative
channels that lead to H-atom photofragments ͓channels ͑3͒,
24
K.-D. Rinnen, M. A. Buntine, D. A. V. Kliner, R. N. Zare, and W. M. Huo,
J. Chem. Phys. 95, 214 ͑1991͒.
W. M. Huo, K.-D. Rinnen, and R. N. Zare, J. Chem. Phys. 95, 205 ͑1991͒.
A. J. R. Heck, W. M. Huo, R. N. Zare, and D. W. Chandler, J. Mol.
Spectrosc. 173, 452 ͑1995͒.
R. N. Bracewell, The Fourier Transform and its Applications ͑McGraw-
Hill, New York, 1986͒.
R. N. Zare, Mol. Photochem. 4, 1 ͑1972͒.
B. Bezard, P. N. Romani, B. J. Conrath, and W. C. Macguire, J. Geophys.
Res. 96, 18961 ͑1991͒.
V. Galasso, Chem. Phys. 161, 189 ͑1992͒.
C. W. Bauschlicher, Jr., K. Haber, H. F. Schaeffer III, and C. F. Bender, J.
Am. Chem. Soc. 99, 3610 ͑1977͒.
M. Sironi, D. L. Cooper, J. Gerratt, and M. Raimondi, J. Am. Chem. Soc.
112, 5054 ͑1990͒.
2
2
5
6
2
7
͑4͒, ͑5͒͒. We estimate, from extrapolation of our limited
2
8
H (v,J) data set, that in the formation of H photofragments,
2
2
29
channel ͑2͒, i.e., the formation of H ϩCH , is somewhat
2
2
3
3
0
1
more important, by a factor of approximately 2, than channel
5͒, i.e., the formation of H ϩCHϩH. If we combine these
͑
2
results and assume no contributions from channel ͑3͒ and
3
2
͑
͑
͑
4͒, the formation of 1,3CH ϩHϩH, we estimate that channel
1͒ is approximately three times more important than channel
2͒, and six times more important than channel ͑5͒. To the
2
3
3
3
4
W. Kirmse, Carbene Chemistry, 2nd ed. ͑Academic, New York, 1971͒.
M. Aoyagi, R. Shephard, A. F. Wagner, T. H. J. Dunning, and F. B. Brown,
J. Phys. Chem. 94, 3236 ͑1990͒.
M. R. Berman and M. C. Lin, J. Chem. Phys. 81, 5743 ͑1984͒.
S. Zabarnick, J. W. Fleming, and M. C. Lin, J. Chem. Phys. 85, 4373
͑1986͒.
extent that the channels ͑3͒ and/or ͑4͒ contribute to the ob-
served slow H atoms ͑13% of the total͒ the relative contri-
butions of channel ͑5͒ and ͑2͒ will diminish.
3
5
36
J. Chem. Phys., Vol. 104, No. 11, 15 March 1996
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
128.59.222.12 On: Mon, 01 Dec 2014 00:06:20